Crew management within South African freight rail

Abstract

Optimisation of resource allocation is a common problem in any operational environment. This research project aims to investigate an alternative method of matching train drivers or crew to trains. Traditionally, crew are matched to trains based on defined train schedules, however this research project tested the impact of a heuristic that matches crew to trains based on random train departures. The impact is measured on the number of successful train crew combinations and the number of loads delivered. Transnet Freight Rail Natal corridor will be used to understand rail operations and crew scheduling to develop a simulation model that increases crew coverage for random train departures. Currently, trains do not depart on schedule and there is significant variability between the train plan and the actual train departures. This is due to many factors that include, but are not limited to infrastructure and rolling stock failures, resulting in misalignment between actual train arrival/departure times and the crew schedules. The impact of such misalignments could lead to any of the following, overtime incurred by crew, trains wait for relief crew to arrive or train cancellations. A simulation model is used to test a proposed future situation in which crew are matched to trains without the use of a schedule or train plan, as previous researchers have done. The model was developed with an embedded heuristic to improve crew coverage over 24 hours. The simulation illustrated an improvement in the number of successful train crew combinations and the number of loads delivered, while reducing variability of crew waiting times and overtime. These findings could contribute to efficiency improvements for the train system and increase asset utilisation.